Development of Activity-Based Probes for Trypsin-Family Serine Proteases Zhengying Pan,* Douglas A

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2882–2885

Development of activity-based probes for trypsin-family serine proteases Zhengying Pan,* Douglas A. Jeﬀery, Kareem Chehade, Jerlyn Beltman,à James M. Clark, Paul Grothaus, Matthew Bogyo*,§ and Amos Baruch* Celera Genomics, 180 Kimball Way, South San Francisco, CA 94080, USA Received 2 February 2006; revised 28 February 2006; accepted 2 March 2006 Available online 22 March 2006

Abstract—A series of diphenylphosphonate-based probes were developed for the trypsin-like serine proteases. These probes selectively target serine proteases rather than general serine hydrolases that are targets for fluorophosphonate-based probes. This increased selectivity allows detection of low abundance serine proteases in complex proteomes using simple SDS–PAGE methods. We present here the application of multiple probes in enzyme activity profiling of intact mast cells, a type of inflammatory cell implicated in allergy and autoimmune diseases. Ó 2006 Elsevier Ltd. All rights reserved.

The primary goal of proteomics is to assign functions to expanded with the development of chemical probes that all proteins in a given cell, tissue or organism.1 The chal- target additional important enzyme families including lenges implicit in this field have generated novel phosphatases5 and kinases.6 Many approaches have approaches to functionally dissect the proteome. Chem- focused on broad-spectrum probes that target multiple ical proteomics or activity-based protein profiling related enzyme family members.7 Thus, there is a great (ABPP) makes use of small molecule probes that form need to develop new chemical probes to selectively covalent complexes with their targets using an activity- target sub-classes of enzymes to study their roles in bio- dependent chemical reaction. These activity-based logical processes. probes (ABPs) can be used to profile enzymatic activities in complex proteomes and provide information on pro- Serine hydrolases represent a large family of enzymes, tein targets at the functional, rather than the expression members of which participate in many crucial biologi- level.2 Moreover, due to its use of small molecules, this cal processes. One of the most intriguing sub-classes of approach naturally focuses on ‘druggable’ enzymes that this family is the trypsin-like serine proteases. This sub- show specific ligand binding. group is comprised of 65 enzymes in humans with un- ique cellular and physiological regulatory roles in A number of classes of ABPs have successfully been health and disease.8 Several enzymes within this group, designed and applied to serine hydrolases3 and cysteine such as thrombin, factor VIIa, factor Xa, and tryptase, proteases.4 Recently the scope of this approach has are being extensively pursued as drug targets in cardiovascular and inflammatory indications. A fluorophosphonate (FP) probe, similar to probe 1 (Bio- Keywords: Activity-based probes; Tryptase; Chemical proteomics. FP), has been previously reported to target the serine * Corresponding authors; e-mail addresses: [email protected]; hydrolase family (Ref. 3). The broad reactivity of [email protected]; [email protected]; amos.baruch@ probe 1 enables simultaneous labeling of a large num- celera.com ber of enzymes including esterases, proteases, lipases, Present address: Novartis Institute of Biomedical Research, Boston, MA 02139, USA. and amidases. In many cases, this probe generates a à Present address: Chiron Corporation, Emeryville, CA 94608, USA. highly complex activity profile that prevents its usage § Present address: Department of Pathology, School of Medicine, for studying low-abundance, specific enzymes using Stanford University, 300 Pasteur Drive, Stanford, CA 94305, USA. simple analytical methods such as SDS–PAGE. To

O NH O O O biotin. Following addition of an amino acid residue at H O O P O HN N O N F the P2 position, similar procedures produced biotinyla- H S H 5 H ted probes 4 and 5. Figure 1. A general serine hydrolase probe 1. Apparent inhibition constants [Ki(app)] were obtained for these probes against four trypsin-like serine proteases circumvent this issue, we developed a series of selective (Table 1). The addition of a P2 proline residue increased probes that specifically target the trypsin-like serine the overall reactivity of probe 4 toward trypsin-like proteases (see Fig. 1). serine proteases. Incorporation of an asparagine residue at P2 yielded a 23-fold selectivity increase for probe 5 To achieve selectivity toward serine proteases, we fo- favoring tryptase over its closely related family member cused our efforts on phosphonate-based probes. Ini- trypsin. tial attempts to use an a-amino fluorophosphonate reactive group failed due to its short half-life in an To confirm that the potency and selectivity observed in aqueous environment.9 The more stable reactive kinetic assays were reflected in enzyme labeling profiles, group, diphenylphosphonate, has previously been probes 1, 4, and 5 were used for activity-based labeling used to generate potent irreversible serine protease of a panel of recombinant enzymes from different serine inhibitors10. Here, we describe our efforts in develop- hydrolase families (Fig. 2). The extent of covalent labeling diphenylphosphonate (DPP)-derived probes for ing was determined by Western blotting with streptavi- activity profiling of trypsin-like serine proteases11 din detection of the biotin reporter group. Probe 1 (see Scheme 1). labeled all recombinant enzymes tested including butyr- ylcholine esterase (BCE), a hydrolase enzyme. In con- Based on previous studies of substrate specificity,12 a trast, the lysine-DPP-based probes were completely lysine residue was incorporated at the P1 site with either inactive against both chymotrypsin and BCE. The gen- a proline or asparagine at the P2 position to generate a eral trypsin-family probe 4 efficiently labeled all trypsin- general trypsin-family protease probe 4 (Bio-PK-DPP) and a b-tryptase-selective probe 5 (Bio-NK-DPP), respectively. Lysine-based DPPs can be made through Table 1. Apparent inhibition constants [Ki(app) lM] of probes 1, 4, and 5 for trypsin-like serine proteases following 30 min of incubation13 the Oleksyszyn 3-component reaction using either con- ventional heating or microwave-assisted agitation. Inter- Enzymes Bio-FP 1 Bio-PK-DPP 4 Bio-NK-DPP 5 mediate 2 was obtained through three steps of b-Tryptase 30 6.2 2.5 protecting group manipulations. A biotinylated probe Trypsin 16.5 0.57 57.3 3 (Bio-K-DPP) was made by coupling 2 with a commer- Thrombin 7.9 1.07 >100 Plasmin >100 2.9 4.03 cially available biotinylation reagent NHS-(PEG)4-

O H O N O O N P O O O a O b

P(OPh)3 O N O O NH2 O O HN H NH H H H O N O O N P O S O O O O O 3 NH2 c O H O H2N P O O O N P O d; c NHH N O O O O O HN N O O O H S H

HN O e; c 4 NH2 O H2N 2 O O NH O O H O H O O N P O HN N O O N O H S H H O 5 NH2

Scheme 1. Synthesis of ABPs with the diphenylphosphonate reactive group. Reagents and conditions: (a) HOAC as solvent, microwave, 150 °C, 5–10 min or oil bath, 70 °C, 1–3 h; (b) hydrazine; Boc2O; H2, Pd–C; (c) NHS-(PEG)4-biotin, triethylamine; then TFA; (d) Cbz-Pro-OH, EDCI, HOBt; H2, Pd–C; (e) Cbz-Asn(Trt)-OH, EDCI, HOBt; H2, Pd–C. 2884 Z. Pan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2882–2885 5 5 4

enzymes were serine hydrolases that were not proteases. P P P P P P P P D

D In addition, several identiﬁed serine hydrolases were D D - 1 - - 1 -

K K P K K P abundant housekeeping enzymes such as acyl-transfer- F P N P N F ------o o o o o o i i i i i i ase, lysophospholipase, and acyl-CoA thioester hydro- B B B B B B lase. In sharp contrast, the trypsin-family-selective β-Tryptase Plasmin probe 3 selectively labeled b-tryptase (Fig. 3, right panel). Further analysis of the labeling with probe 1 re- Trypsin Chymotrypsin vealed that acyl-transferase and b-tryptase co-migrate on the gel, thus making it difficult to distinguish activity-profiles for these two enzymes in HMC-1 cells. Since Thrombin BC Esterase no labeling of acyl-transferase was observed with probe 3, b-tryptase activity could be directly evaluated in the Figure 2. Labeling of recombinant enzymes with probes Bio-FP 1, mast cell proteome without a need for any further Bio-PK-DPP 4, and Bio-NK-DPP 5. purification steps. This exemplifies the utility of protease-selective probes for profiling trypsin-like proteases like enzymes tested. Interestingly, this probe was even within a complex proteome. more potent in labeling thrombin than probe 1 even though it has a much less reactive functional group. HMC-1 cells are immature cells that do not fully reflect Probe 5 labeled three of the four trypsin-like serine pro- properties of mature mast cells. A prominent feature of teases but not thrombin, with a clear preference for b- mast cell is the presence of granules that store pre- tryptase. Together, these results suggest that the incor- formed mediators including proteases such as tryptases. poration of a P2 element enhances both the selectivity Upon activation, the content of the granules is released and potency of the probes for trypsin-like enzymes. into the extracellular milieu to promote local wound healing and inflammatory responses. We characterized Next we determined whether the probes can selectively serine hydrolase and protease activities in more physio- label trypsin-like serine proteases in a complex prote- logically relevant, mature mast cells that were derived ome. Initially, the labeling pattern of the serine protease from CD34+ hematopoietic progenitors. These cells probe 3 was compared to that of the general serine were stimulated to degranulate via activation of the high hydrolase probe 1. Activity-based profiling was per- affinity receptor for IgE (FceR) (Fig. 4). Following acti- formed using the human mast cell line HMC-1. Mast vation, intact cells were incubated with probes for 1 h, cells are inflammatory cells that provide protection conditioned medium was collected, and labeled enzymes against parasitic infections but are also known to play were purified and identified. Probe 1 yielded a complex a major role in allergic responses and autoimmune dis- labeling pattern regardless of the mast cell activation eases.14 Intact HMC-1 cells were incubated for 1 h with state (Fig. 4, lanes 1 and 2). The tryptase-selective probe probe 1 or probe 3. Cell lysates were then analyzed by 5 predominantly labeled b-tryptase (Fig. 4, lanes 3 and SDS–PAGE followed by Western blotting using strepta- 4). Notably, when the general trypsin-family probe 4 vidin-HRP detection reagent (Fig. 3). Multiple active was used for labeling, both tryptase and cathepsin G mast cell serine hydrolases were labeled by probe 1 were labeled (Fig. 4, lanes 5 and 6). The observed reac- (Fig. 3, left panel). Pretreatment with 4-(2-aminoeth- tivity of probe 4 is consistent with the known substrate yl)-benzenesulfonyl fluoride (AEBSF), a general serine specificity of these two enzymes, with tryptase exhibiting hydrolase inhibitor, blocked the labeling of several en- trypsin-like activity and cathepsin G having both tryptic zymes. Affinity purification of probe 1 modified enzymes and chymotryptic activities. Labeling with probes 4 and followed by tryptic digestion and analysis by mass spec- 5 showed clear up-regulation in the activity of these en- trometry, resulted in identification of 15 membrane- zymes following IgE stimulation (Fig. 4, lane 3 vs lane 4 bound and secreted enzymes.15 The majority of these and lane 5 vs lane 6). This demonstrates that probes 4

Bio-FP Bio-NK Bio-PK Bio-FP Bio-K-DPP 1 5 4 1 3 kDa - + - + - + IgE treatment AEBSF - + - + 64 prolyl endopeptidase 51 DPP7 +prolyl carboxypeptidase Acyl-CoA thioester hydrolase 39 β Tryptase 28 Cathepsin G β-tryptase+ acyl transferase RA inducible serine carboxypeptidase 1 2 3 4 5 6 lysophospholipase Figure 4. Selective activity-based labeling of mast cell trypsin-like serine proteases by DPP-based ABPs. Mature human mast cells Figure 3. Selective labeling of b-tryptase in HMC-1 cell lysates by Bio- derived from CD34+ hematopoietic stem cells were induced for K-DPP probe 3, with the absence ( ) and presence (+) of AEBSF degranulation with IgE followed by anti-IgE treatment (lanes 2, 4, À pretreatment prior to probes addition. and 6). Z. Pan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2882–2885 2885 and 5 can be used to study the cellular b-tryptase’s 8. Ossovskaya, V. S.; Bunnett, N. W. Physiol. Rev. 2004, 84, activity upon biological stimulation. 579; Pawlinski, R.; Mackman, N. Crit. Care Med. 2004, 32, S293. In summary, we have demonstrated the design, synthesis, 9. Internal data; see also Ni, L.-M.; Powers, J. C. Bioorg. and application of selective trypsin-family protease-selec- Med. Chem. 1998, 6, 1767. 10. Powers, J. C.; Asgian, J. L.; Ekici, O. D.; James, K. E. tive probes 3, 4, and 5. These probes are capable of specif- Chem. Rev. 2002, 102, 4639. ically labeling trypsin-like serine proteases either in their 11. Hawthorne, S.; Hamilton, R.; Walker, B. J.; Walker, B. pure forms or as components of complex proteomes. Anal. Biochem. 2004, 326, 273. Thus, these probes represent powerful biochemical tools 12. Harris, J. L.; Niles, A.; Burdick, K.; Maffitt, M.; Backes, for monitoring the activity of proteases in their natural B. J.; Ellman, J. A.; Kuntz, I.; Haak-Frendscho, M.; environment. Attachment of other reporter groups such Craik, C. S. J. Biol. Chem. 2001, 276, 34941; Jackson, D. as fluorophores may also allow imaging of protease activ- S.; Fraser, S. A.; Ni, L. M.; Kam, C. M.; Winkler, U.; ity in situ as recently described for cysteine proteases.16 Johnson, D. A.; Froelich, C. J.; Hudig, D.; Powers, J. C. Finally, probes that target other sub-families of serine J. Med. Chem. 1998, 41, 2289. proteases can be designed using a similar approach. 13. Inhibition profiles for the serine proteases were generated by incubating each enzyme in the presence of probes Increasing the number and type of chemical proteomics (various concentrations) or 10% DMSO (vehicle control) probes targeting serine proteases will significantly en- for 30 min in 96-well clear polystyrene plates at room hance our ability to understand the activity, regulation, temperature prior to the addition of substrate. Enzyme and function of these important enzymes. activity was measured by monitoring the hydrolysis of the synthetic substrate tosyl-Gly-Pro-Lys-para-nitroanilide at 405 nM over 5 min using a UV/MAX kinetic plate reader References and notes (Molecular Devices). The apparent inhibition constants [Ki(app)] were calculated from the progress curves using the 1. Saghatelian, A.; Cravatt, B. Nat. Chem. Biol. 2005, 1, 130. software package Batch Ki (BioKin, Ltd). In each case, the 2. Jeffery, D. A.; Bogyo, M. Curr. Opin. Biotechnol. 2003, 14, substrate concentration was at or below the Km for the 87; Adam, G. C.; Cravett, B. F.; Sorensen, E. J. Nat. enzyme. Biotechnol. 2002, 20, 805. 14. Benoist, C.; Mathis, D. Nature 2002, 420, 875. 3. Liu, Y.; Patricelli, M. P.; Cravatt, B. F. Proc. Natl. Acad. 15. The following hydrolases were identified by LC/MS/MS Sci. U.S.A. 1999, 96, 14694. analysis (Q-Star) (serine protease in bold). Secreted 4. Greenbaum, D.; Medzihradszky, K. F.; Burlingame, A.; serine hydrolases: dipeptidyl peptidase 7 (54 kDa); Bogyo, M. Chem. Biol. 2000, 7, 569. esterase 10 (37 kDa); tryptase b1 (34 kDa); lysophos- 5. Kumar, S.; Zhou, B.; Liang, F.; Wang, W.-Q.; Huang, Z.; pholipase 2 (24 kDa); platelet-activating factor acetyl- Zhang, Z.-Y. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, hydrolase (26 kDa). Membrane serine hydrolases: 7943. oxidized protein hydrolase (82 kDa); prolyl endopepti- 6. Liu, Y.; Shreder, K. R.; Gai, W.; Corral, S.; Ferris, D. K.; dase (81 kDa); angiotensinase C/propylcarboxypeptidase Rosenblum, J. S. Chem. Biol. 2005, 12, 99; Yee, M.-C.; (56 kDa); retinoid-inducible serine carboxypeptidase Fas, S. C.; Stohlmeyer, M. M.; Wandless, T. J.; Cimprich, (51 kDa); KIAA1363/carboxy esterase (48 kDa); perox- K. A. J. Biol. Chem. 2005, 280, 29053; Cohen, M. S.; isomal acyl-coenzyme A thioesterhydrolase 2a (46 kDa); Zhang, C.; Shokat, K.; Taunton, J. Science 2005, 308, aldolase A (39 kDa); protein FLJ11342/lipase (32 kDa); 1318. similar to hypothetical protein 4833421E05Rik/lipase 7. Greenbaum, D. C.; Baruch, A.; Grainger, M.; Bozdech, (28kDa); OVCA2/tumor suppressor in ovarian cancer 2 Z.; Medzihradszky, K. F.; Engel, J.; DeRisi, J.; Holder, A. (25 kDa). A.; Bogyo, M. Science 2002, 298, 2002; Jessani, N.; Liu, 16. Joyce, J. A.; Baruch, A.; Chehade, K.; Meyer-Morse, N.; Y.; Humphrey, M.; Cravatt, B. F. Proc. Natl. Acad. Sci. Giraudo, E.; Tsai, F. Y.; Greenbaum, D. C.; Hager, J. H.; U.S.A. 2002, 99, 10335. Bogyo, M.; Hanahan, D. Cancer Cell 2004, 5, 443.